Gagosian’s DALL-E–Enabled Art Exhibition Throws Us Headfirst into the Uncanny Valley

The arrival of AI text generators and chatbots like Chat GPT and Bing (or is she named Sydney?) over the last year has shattered the assumption that creativity is the sole domain of humans, and other living things. But, while image generators like DALL-E and Midjourney are the visual equivalent technologies, the same crisis has not quite registered in the art world.

Perhaps, this lack of response stems from a lack of opportunity. No longer! Earlier this week, mega-gallery Gagosian opened an exhibition of works by DALL-E, which, like its AI image generator competitors, can turn a simple text prompt into an image in seconds. Might I find some crisis awaiting me there? (Yes).

The exhibition is produced by Bennett Miller, a film director who has been nominated for Oscars for Foxcatcher (2014) and Capote (2005); the works, and the exhibition are untitled. Over the past several years, Miller has been making a documentary about AI, through which he interviewed Sam Altman, the CEO of OpenAI, who gave him beta access to DALL-E far before the rest of the public.

The images DALL-E produces range from obviously amiss (twisted fingers, a fuzzy swirl of pixels) to hauntingly accurate in their targeting of one’s request. Despite these occasional flaws, no longer is the AI image quickly clocked for what it is by that tell-tale sheen of psychedelic patterning. It’s no wonder then why the word “real” was invoked, again and again, by the audience at Miller’s opening this week.

One woman I pass gestures at one of Miller’s prints, a large piece laid on with deep, dark, wet-looking ink onto sepia-toned paper, depicting a child as she stares at the viewer while the wind tosses her hair. It looks as if it comes from the Victorian era, dated not just by its coloring but by what looks to be a simple, linen dress of the era. It’s all projection. The woman tells her friend, “It’s not real.” There is no linen dress.

Well, so what. It’s a bit melodramatic to behave as if we don’t already live in an era of unreal-ness. And anyways, since when does art require a real-world referent to represent something “real”? Since when is “realness” a metric?

Sure, many of Miller’s works look like they could be photographs, but many are heavily stylized. Often extremely out of focus and piled on with grain, there is just enough form to suggest a subject or a landscape. Some of them seem to represent momentous or historical moments in the past. Here is a profile that looks Native American, extending an arm that could be a wing, that could be cultural dress. Here is a mushroom cloud, as if from an explosion, but flattened in a way that, perhaps, Nature wouldn’t allow. A machine like a train but it’s not. A disk, just a flat circle of some substance, held in the hands of a woman. Beguilingly simple, pointing back to nothing.

I spot Fran Lebowitz. Blunt, coarse bob, big coat, tortoiseshell glasses perched on her nose and another set in her welt pocket. Loafers! It really is her. She’s thumbing through the exhibition text that was produced for the show by author Benjamin Labatut partially using ChatGPT, an AI text generator also produced by OpenAI. It turns out Miller also interviewed Lebowitz for his documentary, though it doesn’t seem clear why. She repeats an apology to me several times: she doesn’t know what this means, the exhibition, the fact of its genesis. But she makes an effort.

“These are not real photographs, but what are real photographs?” Lebowtiz begins. “Are the only real photographs the ones made on film, not the digital ones? My friend Peter Hujar would say so.”

The slippery slope tack: if we’ve accepted that cameras do not make the photographs, but that photographers do, why should any succeeding technology that the human mind directs for its purpose not be judged similarly? That is, as a genuine, human act of creation. I ask Lebowitz a clumsy question, something like, ‘Isn’t the labor of trying to make something worth something?” She says of course. What are we even talking about? It’s too basic but I can’t help it.

The concern about realness comes from two places. Where did these images come from and can we credit Miller with a “real” creative act. It’s really one problem: what do we do with this other actor in the picture, AI? What spasm was it that gave birth to these images, that Miller guided and curated?

It’s telling that these new tools are called AI “generators” not “creators”. Generation is to bring into being, but behind a veil. Generation has its roots in the phenomenon of conception, which is not done with the conscious mind but the secret efforts of the body. It is only in this way that I can relate to the concept of AI, this thing that brings into being without conscious, all the indifference and capability of nature. But this is false analogy (is there a word for anthropomorphizing but for nature? Naturmorphizing?). I’m not sure why I can’t see it as an extension of all the other amazing technological capabilities with their hidden mechanisms. I don’t know how my computer works.

Walking around Miller’s show I’m surprised that so many people look happy and curious whereas I feel bitterly on guard. I look closely at each image, which range from looking like vintage photographs to charcoal drawings, and investigate for signs of their computerly origins. I’m not to be tricked!

As images, though, I do like them. They remind me of a picture book I once had and spark my love of old and whimsical looking things, for what that’s worth. A lot of AI images I’ve seen do this, that is, open the door to alternate, fantastical worlds, which says a lot about the people who request these images. There’s a lovely impulse to see something wondrous, magical, not of our reality. But how tightly and terribly joined is this desire for the fantastic to the impish twitch for falsity.

By now, haven’t we all seen those AI generated images of Trump getting arrested? How quickly we come back to Earth. One day it’ll feel normal. For now it’s tripping me up.